Vehicle control method, vehicle control device, computer device, and storage medium

Abstract

The present invention discloses a vehicle control method, a vehicle control device, a computer device, and a computer storage medium. Through the vehicle control method, the vehicle control parameters are adjusted in a timely manner according to the driver's psychological data, and the vehicle operation is controlled according to the adjusted control parameters. The vehicle control method includes: determining first psychological data of the driver corresponding to the target vehicle, where the first psychological data includes emotion data and / or attention data; inputting the first psychological data into a control parameter adjustment model to obtain a first calculated control parameter; and adjusting the target vehicle based on the first calculated control parameter.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving, and particularly to a method for controlling an autonomous driving vehicle and related devices. Background Art

[0002] In recent years, the development of autonomous driving has, to a certain extent, reduced the labor intensity of drivers in driving vehicles and effectively reduced the fatigue caused by drivers' long-time control of vehicles.

[0003] The focus of the control of existing autonomous driving vehicles is mainly on the vehicle itself, without taking the feelings or emotions brought by the system to users as control factors and inputting them into the vehicle for closed-loop control. For example, once the following distance of an Adaptive Cruise Control (ACC) is set, the system will follow the vehicle according to the set parameters. However, during the following process, if the user feels that the following distance is too close and panics, the system doesn't know and doesn't take the panic as a system control input to adjust the system parameters. As a result, it may bring discomfort or distrust to users in using the autonomous driving system. Summary of the Invention

[0004] In order to solve the technical problem in the prior art that the feelings or emotions brought by the system to users are not used as control inputs for closed-loop control, the present invention provides a method and related devices for actively identifying the driver's emotions based on the driver's physiological parameters and adjusting the control parameters of the autonomous driving vehicle based on the driver's emotions.

[0005] The first aspect of the present invention provides a vehicle control method, including the following steps;

[0006] Determine the first psychological data of the driver corresponding to the target vehicle, where the first psychological data includes emotion data and / or attention data;

[0007] Input the first psychological data into a control parameter adjustment model to obtain a first calculated control parameter. The control parameter adjustment model is obtained by training control parameter training samples, and the control parameter training samples include training psychological data corresponding to at least one driver and control parameters corresponding to the training psychological data;

[0008] Adjust the target vehicle based on the first calculated control parameter.

[0009] Optionally, the determining the first psychological data of the driver corresponding to the target vehicle includes the following steps:

[0010] Collect the first body characteristic parameters of the driver;

[0011] Input the first body characteristic parameters into a driver recognition model to obtain the first psychological data.

[0012] Optionally, the method further includes the following steps:

[0013] Obtain the physical characteristic parameters corresponding to at least one driver and the actual psychological data corresponding to the physical characteristic parameters;

[0014] Preprocess the physical characteristic parameters corresponding to the at least one driver and the actual psychological data corresponding to the physical characteristic parameters;

[0015] Iteratively run according to the preprocessed data and the initial driver recognition model until a preset iteration termination condition is reached;

[0016] Determine the initial driver recognition model when the preset iteration termination condition is reached as the driver recognition model.

[0017] Optionally, the method further includes the following steps:

[0018] Judge whether the number of iterations reaches a preset value. If so, determine that the preset iteration termination condition is satisfied;

[0019] Or,

[0020] Judge whether the model parameters corresponding to the initial driver recognition model converge. If so, determine that the preset iteration termination condition is satisfied.

[0021] Optionally, the method further includes the following steps:

[0022] Determine the second psychological data of the driver after adjusting the target vehicle based on the first calculation control parameter;

[0023] If the second psychological data does not reach a preset adjustment threshold, determine a second calculation control parameter based on the second psychological data and the control parameter adjustment model;

[0024] Adjust the target vehicle based on the second calculation control parameter.

[0025] Optionally, the method further includes the following steps:

[0026] Determine the expected return function corresponding to the control parameter adjustment model;

[0027] Evaluate the control parameter output by the control parameter adjustment model through the expected return function;

[0028] Update the control parameter adjustment model according to the evaluation result.

[0029] Optionally, the method further includes the following steps:

[0030] Determine whether the second calculation control parameter reaches a preset parameter threshold;

[0031] If the second calculation control parameter reaches the preset parameter threshold, send a prompt message;

[0032] Adjust the control parameter according to the corresponding feedback information sent by the driver in response to the prompt message;

[0033] Adjust the target vehicle according to the adjusted control parameter.

[0034] The second aspect of the present invention provides a vehicle control device, including:

[0035] A determination unit, configured to determine first psychological data of a driver corresponding to a target vehicle, where the first psychological data includes emotion data and / or attention data;

[0036] A calculation unit, configured to input the first psychological data into a control parameter adjustment model to obtain a first calculation control parameter, where the control parameter adjustment model is obtained by training a training sample, and the training sample includes training psychological data corresponding to at least one driver and control parameters corresponding to the training psychological data;

[0037] An adjustment unit, configured to adjust the target vehicle based on the first calculation control parameter.

[0038] Optionally, the determination unit is specifically configured to:

[0039] Collect first physical characteristic parameters of the driver;

[0040] Input the first physical characteristic parameters into a driver recognition model to obtain the first psychological data.

[0041] Optionally, the determination unit determines the driver recognition model through the following steps:

[0042] Obtain physical characteristic parameters corresponding to at least one driver and actual psychological data corresponding to the physical characteristic parameters;

[0043] Preprocess the physical characteristic parameters corresponding to the at least one driver and the actual psychological data corresponding to the physical characteristic parameters;

[0044] Iteratively run according to the preprocessed data and an initial driver recognition model until a preset iteration termination condition is reached;

[0045] Determine the initial driver recognition model when the preset iteration termination condition is reached as the driver recognition model.

[0046] Optionally, the determining unit determines the iteration termination condition through the following steps:

[0047] Judge whether the number of iterations reaches a preset value. If so, it is determined that the preset iteration termination condition is satisfied;

[0048] Or,

[0049] Judge whether the model parameters corresponding to the initial driver recognition model converge. If so, it is determined that the preset iteration termination condition is satisfied.

[0050] Optionally, the determining unit is further configured to:

[0051] Determine the second psychological data of the driver after adjusting the target vehicle based on the first calculation control parameter;

[0052] If the second psychological data does not reach the preset adjustment threshold, determine the second calculation control parameter based on the second psychological data and the control parameter adjustment model;

[0053] Adjust the target vehicle based on the second calculation control parameter.

[0054] Optionally, the vehicle control device further includes:

[0055] An updating unit, configured to:

[0056] Determine the expected return function corresponding to the control parameter adjustment model;

[0057] Evaluate the control parameter output by the control parameter adjustment model through the expected return function;

[0058] Update the control parameter adjustment model according to the evaluation result.

[0059] Optionally, the adjusting unit is further configured to:

[0060] Judge whether the second calculation control parameter reaches a preset parameter threshold;

[0061] If the second calculation control parameter reaches the preset parameter threshold, send a prompt message;

[0062] Adjust the control parameter according to the corresponding feedback information sent by the driver for the prompt message;

[0063] Adjust the target vehicle according to the adjusted control parameter.

[0064] The third aspect of the present invention provides a computer device, including: at least one connected processor, a memory, and a transceiver; wherein, the memory is used for storing program codes, and the processor is used for calling the program codes in the memory to execute the steps of the vehicle control method described in the first aspect above.

[0065] The fourth aspect of the embodiments of the present invention provides a computer storage medium, which includes instructions that, when running on a computer, cause the computer to execute the steps of the vehicle control method described in the first aspect above.

[0066] Compared with the related art, in the embodiments provided by the present invention, the vehicle control device determines the first psychological data of the target vehicle driver, determines the first calculation control parameter according to the first psychological data, and adjusts the running state of the target vehicle according to the first calculation control parameter, so that the vehicle control parameter can be adjusted in a timely manner according to the driver's psychological data, and the vehicle operation can be controlled according to the adjusted control parameter, providing a better autonomous driving experience for users. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 It is a system architecture diagram of the vehicle control device provided by the embodiments of the present invention;

[0068] Figure 2 It is a flow schematic diagram of the vehicle control method provided by the embodiments of the present invention;

[0069] Figure 3 It is a flow schematic diagram of the driver recognition model training method provided by the embodiments of the present invention;

[0070] Figure 4 It is a training flow schematic diagram of the driver recognition model training provided by the embodiments of the present invention;

[0071] Figure 5 It is a virtual structure diagram of the vehicle control device provided by the embodiments of the present invention;

[0072] Figure 6 It is a hardware structure diagram of the server provided by the embodiments of the present invention; and

[0073] Figure 7 It is a dynamic optimization schematic diagram of the control parameter adjustment model. DETAILED DESCRIPTION OF THE EMBODIMENTS [[ID=...]]

[0074] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments.

[0075] The present invention provides a vehicle control method, which is a vehicle control method based on active human-computer interaction. It takes the driver's psychological data as input, the vehicle control parameters as output, and adjusts the vehicle based on the output control parameters. This method can feedback the driver's subjective factors to the input of the autonomous driving system and change the vehicle control parameters to improve the driver's driving experience.

[0076] Please refer to Figure 1 , which is the system architecture diagram of the vehicle control device provided by the embodiment of the present invention.

[0077] In this embodiment, the system architecture of the vehicle control device includes a driver detection sensor 101, a vehicle status sensor 102, a vehicle control device 103, and a vehicle 104.

[0078] The driver detection sensor 101 is used to collect the driver's body characteristic parameters and send the driver's body characteristic parameters to the vehicle control device 103. Among them, the driver detection sensor 101 can include an in-vehicle camera, or can include in-vehicle millimeter-wave radar and other devices installed on the vehicle, or can also include a bracelet or other wearable devices that are Bluetooth-connected to the vehicle 104 and are not fixedly installed on the vehicle 104. Of course, it can also include multiple devices at the same time. For example, there is both an in-vehicle camera and a wearable device Bluetooth-connected to the vehicle. The specific is not limited. The body characteristic parameters collected by the driver detection sensor 101 include face data (such as facial expressions, eye spatial positions, blink frequencies, pupil focus positions, etc., the specific is not limited), head data (such as breathing frequencies, etc., the specific is not limited), hand data (such as hand postures and grip strengths, etc., the specific is not limited), heart rate data, etc. Of course, it can also include leg postures (such as whether to adopt a pre-braking posture, etc.), the specific is not limited.

[0079] The vehicle sensor 102 collects vehicle status data and surrounding environment data, and sends the collected vehicle status data and surrounding environment data to the vehicle control device 103. Among them, the vehicle status data includes vehicle speed, following distance, sensitivity parameters of lane departure warning (LDW), etc., which are not specifically limited; the environmental data includes the number of lanes of the road on which the vehicle 104 is running, the lane attribute of the vehicle 104, other vehicle information around the vehicle 104 (for example, on a four-lane road, the vehicle 104 is located in the second lane on the left side of the road, the vehicle speed is 90 km / h, the speed limit of this lane is 100 km / h, there is a car with a speed of 80 km / h in front of the vehicle 104, there is a car with a speed of 120 km / h 20 m in front of the left lane of the vehicle 104, and there is a truck with a speed of 120 km / h 30 m in front of the right lane), of course, it can also include the environmental information around the road (for example, there are cliffs or lakes on both sides of the road), which is not specifically limited.

[0080] The vehicle control device 103 receives the driver physiological characteristic parameters collected by the driver sensor 101, as well as the vehicle status data and surrounding environment data collected by the vehicle status sensor. The vehicle control device 103 inputs the received physical characteristic parameters into the driver recognition model, and outputs the psychological data corresponding to the physical characteristic parameters. The psychological data is attention data and / or emotion data. The vehicle control device 103 inputs the psychological data into the control parameter adjustment model, and outputs the vehicle control parameters corresponding to the psychological data. The vehicle control device 103 issues a control instruction according to the output vehicle control parameters to adjust the vehicle 104. Adjusting the vehicle 104 includes adjusting the vehicle speed, adjusting the following distance of the vehicle, and adjusting the alarm threshold of lane departure warning, etc., which are not specifically limited.

[0081] Please refer to Figure 2 , which is a schematic flowchart of the vehicle control method provided by the embodiment of the present invention, including:

[0082] 201. Determine the first psychological data of the driver corresponding to the target vehicle.

[0083] In this embodiment, the vehicle control device can collect the first physical characteristic parameters of the driver corresponding to the target vehicle through the driver detection sensor, input the first physical characteristic parameters into the driver recognition model, and output the first psychological data corresponding to the first physical characteristic. The first psychological data includes emotion data and / or attention data.

[0084] 202. Input the first psychological data into the control parameter adjustment model to obtain the first calculated control parameter.

[0085] In this embodiment, the control parameter adjustment model is obtained by training control parameter training samples, which include training psychological data corresponding to at least one driver and control parameters corresponding to the training psychological data. Among them, the psychological data includes emotion data and / or attention data, and the control parameters include the ACC following distance, vehicle speed, LDW alarm threshold, etc., which are not specifically limited.

[0086] It should be noted that the control parameter adjustment model can be an ACC following distance parameter adjustment model, an LDW alarm threshold parameter adjustment model, or other parameter adjustment models (such as a cruise control vehicle speed adjustment model). Of course, it can also include multiple parameter adjustment models at the same time, inputting a psychological data to output multiple control parameters such as the ACC following distance parameter and the LDW alarm threshold parameter at the same time.

[0087] It can be understood that the control parameters corresponding to the training psychological data can be collected based on the subjective questionnaire method. For example, when the driver believes that in adaptive cruise control, when the following distance is 100m, it belongs to a relaxed emotional state, then the following distance of 100m is the control parameter, and the emotional state of relaxation is the psychological data. The control parameter adjustment model takes the driver's psychological data as the input and the control parameters as the output. The offline training method of the control parameter adjustment model is the same as that of the driver recognition model and will not be elaborated here.

[0088] It should be noted that in the traditional control system, the control target is generally directly given in the form of a given quantity, while in the intelligent control system, the control target is sometimes not clear or cannot be directly obtained. In the present invention, after the driver recognition model calculates and outputs the driver emotion data information, the control parameter adjustment model is used to identify the driver emotion data. The following details the specific method of driver emotion recognition:

[0089] 1. System parameter identification:

[0090] Determine the state describing the system according to the system input and output. For the complex non-linear dynamic system jointly composed of the driver's emotion, action, vehicle, and surrounding environment, it is difficult to describe it using a linear function or based on prior knowledge. In the present invention, a deep neural network (i.e., the control parameter adjustment model) is used for system identification to enable it to have the ability to fit complex non-linear inputs, and the system identification is transformed into the parameter optimization of the deep neural network.

[0091] 2. Optimization of the control target:

[0092] In the present invention, a deep neural network is used to fit the system dynamic model, that is, a control parameter adjustment model is trained by using data over a period of time, which includes driver emotions, actions, vehicle motion states, and the positional relationship between the vehicle and the surrounding environment (such as the change in the distance between the vehicle and the vehicle ahead, the deviation distance of the vehicle from the lane, etc.).

[0093] While predicting the future trend of driver emotions through the trained control parameter adjustment model, the driver emotions, actions, vehicle motion states, and the positional relationship between the vehicle and the surrounding environment (such as the change in the distance between the vehicle and the vehicle ahead, the deviation distance of the vehicle from the lane, etc.) are comprehensively considered. Thus, the future overall system state can be predicted (including changes in driver emotions, actions and action amplitudes, and vehicle motion state changes caused thereby), and the adjustment parameter is determined by using the state change value between the state at the current moment and the state at the next moment predicted (for example, the state of the vehicle such as the following distance or vehicle speed, taking the following distance as an example for illustration, the current following distance is 50m, and the following distance at the next moment is 80m, and the adjustment parameter is to increase the following distance by 30m).

[0094] After the above system dynamic prediction, an expected return function is introduced into the control parameter adjustment model. Thus, the control parameter adjustment model can evaluate the adjustment parameter predicted each time through the expected return function and optimize the control parameter adjustment model according to the evaluation result. The following is combined with Figure 7 for illustration:

[0095] Please refer to Figure 7 , which is a schematic diagram of the dynamic optimization of the control parameter adjustment model. Among them, S0 represents the state of the system at the current moment, S1 represents the predicted state of the system at the next moment, V0 represents the action taken by the system at the current moment, and G represents the expected return function when the system executes the action V0 in the current state S0.

[0096] Thus, the optimal action is executed at each step and the deep neural network is trained according to the return generated at each step, so as to obtain the calculation result for identifying and optimizing the vehicle motion control parameters based on the driver emotions and behaviors as inputs.

[0097] 203. Adjust the target vehicle based on the first calculated control parameter.

[0098] In this embodiment, the vehicle control device issues an instruction to the vehicle according to the first calculated control parameter to adjust the running state of the target vehicle. For example, when the following distance of the target vehicle during running is 50m and the following distance in the first calculated control parameter is 70m, the vehicle control device issues an instruction according to the first calculated control parameter to adjust the vehicle distance of the target vehicle to 70m.

[0099] It should be noted that after the vehicle control device adjusts the target vehicle based on the first calculated control parameter, the following operations are also performed:

[0100] Determine the driver's second psychological data after adjusting the target vehicle based on the first calculated control parameter;

[0101] If the second psychological data does not reach the preset adjustment threshold, determine the second calculated control parameter based on the second psychological data and the control parameter adjustment model;

[0102] Adjust the target vehicle based on the second calculated control parameter.

[0103] In this embodiment, after the vehicle control device adjusts the target vehicle based on the first calculated control parameter, it can also determine the driver's second psychological data, which includes emotion data and / or attention data. Then, it is judged whether the second psychological data reaches the preset adjustment threshold. For example, to judge whether the driver is in a relaxed state, the corresponding tension state of the driver can be graded, such as first-level tension, second-level tension, relaxed state, etc. This grading has an associated relationship with the preset adjustment threshold; if it is determined that the second psychological data does not reach the preset adjustment threshold (for example, the preset adjustment threshold is set to the relaxed state, and the second psychological data is still in the tense state), then determine the second calculated control parameter based on the second psychological data and the control parameter adjustment model; adjust the target vehicle based on the second calculated control parameter.

[0104] It can be understood that after the vehicle control device continuously adjusts the vehicle, it judges whether the second calculated control parameter reaches the preset parameter threshold; if so, a prompt message is sent, and the driver makes a feedback according to the prompt message, and adjusts the control parameter according to the corresponding feedback message sent by the driver for the prompt message; adjust the target vehicle according to the adjusted control parameter. That is, if the control parameter of the vehicle is adjusted multiple times, for example, after increasing the following distance of the vehicle multiple times, the driver is still in a tense state, then a voice prompt message can be sent to interact with the driver, such as asking the driver whether to continue to increase the following distance, or asking the driver whether to stop and rest.

[0105] In addition, after determining the second psychological data and the first calculated control parameter, an update operation can be performed on the control parameter adjustment model through the second psychological data and the first calculated control parameter.

[0106] The following takes the adjustment of the following distance parameter of adaptive cruise (ACC) and the adjustment of the alarm threshold parameter of lane departure warning (LDW) as examples to describe in detail how to adjust the target vehicle according to the calculated control parameter:

[0107] I. ACC following distance adjustment:

[0108] First, offline train a driver recognition model and an ACC following distance parameter adjustment model. Among them, the ACC following distance parameter adjustment model is a type of control parameter adjustment model, which takes the emotional state in the driver's psychological data as the input and the following distance as the output.

[0109] Secondly, the vehicle control device collects the driver's physical characteristic data through the driver sensor, inputs it into the driver recognition model, and outputs the driver's first emotion.

[0110] Thirdly, input the driver's first emotion into the ACC following distance parameter adjustment model, output the first following distance parameter, and adjust the vehicle's following distance to be consistent with the first following distance parameter. If the first emotion is in a tense state, then increase the vehicle's following distance.

[0111] Finally, after adjusting the vehicle's following distance to be consistent with the first following distance parameter, determine the driver's second emotion again, and update the ACC following distance parameter adjustment model according to the second emotion and the first following distance parameter. If the first emotion is in a tense state, determine whether the second emotion has reached a relaxed state. If not, then judge whether the following distance has reached the preset parameter threshold, for example, whether it has reached 250m. If so, send a prompt message asking the driver whether to increase the following distance. If the driver answers yes, adjust the vehicle according to the feedback of the driver to the prompt message and increase the following distance.

[0112] II. LDW alarm threshold parameter adjustment:

[0113] First, offline train a driver recognition model and an LDW alarm threshold parameter adjustment model. Among them, the LDW alarm threshold parameter adjustment model is a type of control parameter adjustment model, which takes the driver's attention as the input and the LDW alarm threshold parameter as the output.

[0114] Secondly, the vehicle control system collects the driver's physical characteristic data through the driver sensor, inputs it into the driver recognition model, and outputs the driver's first attention.

[0115] Thirdly, input the driver's first attention into the LDW alarm threshold parameter adjustment model, output the first alarm parameter, and adjust the vehicle's LDW alarm threshold to be consistent with the first alarm parameter. If the first attention is concentrated, then increase the LDW alarm threshold so that the system alarm will not be too sensitive to interfere with the driver's normal driving; if the first attention is distracted or divided, then decrease the LDW alarm threshold to timely remind the driver that they have deviated from the lane.

[0116] Finally, after adjusting the LDW alarm threshold of the vehicle to be consistent with the first alarm parameter, interact with the driver through voice prompts to ask if the LDW alarm threshold is appropriate. If so, obtain the driver's second attention, and update the LDW alarm threshold parameter adjustment model according to the second attention and the first alarm parameter.

[0117] In summary, in the embodiments provided by the present invention, by determining the first psychological data of the driver of the target vehicle, determining the first calculation control parameter according to the first psychological data, and adjusting the running state of the target vehicle according to the first calculation control parameter, the vehicle control parameters can be adjusted in a timely manner according to the driver's psychological data, and the vehicle operation can be controlled according to the adjusted control parameters, providing a better autonomous driving experience for users.

[0118] Next, in combination with Figure 3 a detailed description of the training method of the driver recognition model will be given. Please refer to Figure 3 , which is a schematic flowchart of the driver recognition model training method provided by the embodiments of the present invention, including:

[0119] 301. Obtain the body feature parameters corresponding to at least one driver and the actual psychological data corresponding to the body feature parameters.

[0120] In this embodiment, the vehicle control device can obtain the body feature parameters corresponding to at least one driver and the actual psychological data corresponding to the body feature parameters from the sample database, and the sample database stores the sample data corresponding to multiple drivers.

[0121] 302. Preprocess the body feature parameters corresponding to at least one driver and the actual psychological data corresponding to the body feature parameters.

[0122] In this embodiment, after the vehicle control device obtains the body feature parameters corresponding to at least one driver and the actual psychological data corresponding to the body feature parameters, it can preprocess them, and the preprocessing includes but is not limited to removing unique attributes, processing missing values, feature encoding, data standardization, and feature selection.

[0123] 303. Iteratively run according to the preprocessed data and the initial driver recognition model until the preset iteration termination condition is reached.

[0124] 304. Determine the initial driver recognition model when the preset iteration termination condition is reached as the driver recognition model.

[0125] In this embodiment, the vehicle control device can iteratively train the preprocessed data through a neural network until a preset iteration termination condition is reached, and determine the initial driver recognition model when the preset iteration termination condition is reached as the driver recognition model.

[0126] It should be noted that the iteration termination condition can be judged through the following steps:

[0127] Judge whether the number of iterations reaches a preset value. If so, it is determined that the preset iteration termination condition is met;

[0128] Or,

[0129] Judge whether the model parameters corresponding to the initial driver recognition model converge. If so, it is determined that the preset iteration termination condition is met.

[0130] Next, in combination with Figure 4 The offline training process of the driver recognition model will be described in detail. Please refer to Figure 4 , which is a schematic diagram of the training process for training the driver recognition model provided in the embodiments of the present invention.

[0131] The training computer device 401 obtains a training sample set from the database 402. The training sample set includes body feature parameters corresponding to at least one driver and actual psychological data corresponding to the body feature parameters. It can be understood that the body feature parameters corresponding to at least one driver include data such as the driver's face, head, hand, heartbeat, etc., and the psychological data corresponding to the body feature parameters include at least one of emotion data and attention data, which can be collected by means of a subjective evaluation questionnaire.

[0132] An initial training model is configured on the training computer device 401. The initial training model is the initial driver recognition model. The training computer device 401 can iteratively train the initial training model through the training sample set to determine the reference model parameters for iteratively updating the initial training model until a preset iteration termination condition is reached; input the model parameters into the initial training model to determine the target training model. After each iteration is completed, judge whether the number of iterations reaches a preset value (for example, the preset number of iterations is 1000 times). If so, it is determined that the preset iteration termination condition is met, and the initial training model at the end of the iteration is determined as the driver recognition model; or, after each iteration is completed, judge whether the model parameters corresponding to the initial training model converge. If so, it is determined that the preset iteration termination condition is met, and the initial training model at the end of the iteration is determined as the driver recognition model.

[0133] It should be noted that the initial training model can be understood as a function of a deep neural network, where the coefficients in the function are in an unknown state. These unknown coefficients can be understood as the model parameters of the initial training model. Each physical feature information can be understood as multiple input parameters, and the corresponding psychological data can be understood as the corresponding output parameters. The relationship between the input parameters and the output parameters can be expressed as y = f(x1, x2,..., xN). By inputting many groups of input parameters and output parameters into the initial training model, these unknown model parameters can be determined, thereby completing the training of the model and obtaining the target training model. The target training model is determined to be a driver recognition model. In this way, after inputting the driver's physical feature parameters corresponding to the target vehicle into the driver recognition model, the first physical feature parameter information (x1, x2,..., xN) of the driver corresponding to the target vehicle can be extracted, and the corresponding output parameter y can be output to obtain the first psychological data corresponding to the first physical feature parameter. The above explains the present invention from the perspective of the vehicle control method. The following will explain it from the perspective of the vehicle control device.

[0134] Please refer to Figure 5 , which is a schematic virtual structure diagram of the vehicle control device provided by an embodiment of the present invention. The vehicle control device 500 includes:

[0135] A determination unit 501, configured to determine the first psychological data of the driver corresponding to the target vehicle, where the first psychological data includes emotion data and / or attention data;

[0136] A calculation unit 502, configured to input the first psychological data into a control parameter adjustment model to obtain a first calculated control parameter. The control parameter adjustment model is obtained by training a training sample, and the training sample includes the training psychological data corresponding to at least one driver and the control parameters corresponding to the training psychological data;

[0137] An adjustment unit 503, configured to adjust the target vehicle based on the first calculated control parameter.

[0138] Optionally, the determination unit 501 is specifically configured to:

[0139] Collect the first physical feature parameters of the driver;

[0140] Input the first physical feature parameters into a driver recognition model to obtain the first psychological data.

[0141] Optionally, the determination unit 501 determines the driver recognition model through the following steps:

[0142] Obtain the physical feature parameters corresponding to at least one driver and the actual psychological data corresponding to the physical feature parameters;

[0143] Preprocessing the physical characteristic parameters corresponding to the at least one driver and the actual psychological data corresponding to the physical characteristic parameters;

[0144] Iterate based on the pre-processed data and the initial driver identification model until a preset iteration termination condition is reached;

[0145] The initial driver recognition model when the preset iteration termination condition is reached is determined as the driver recognition model.

[0146] Optionally, the determining unit 501 determines the iteration termination condition by the following steps:

[0147] Determine whether the number of iterations reaches a preset value, and if so, determine whether the preset iteration termination condition is met;

[0148] or,

[0149] It is determined whether the model parameters corresponding to the initial driver identification model converge, and if so, whether the preset iteration termination condition is satisfied.

[0150] Optionally, the determining unit 501 is further configured to:

[0151] determining second psychological data of the driver after adjusting the target vehicle based on the first calculated control parameter;

[0152] If the second psychological data does not reach a preset adjustment threshold, determining a second calculation control parameter based on the second psychological data and the control parameter adjustment model;

[0153] The target vehicle is adjusted based on the second calculated control parameter.

[0154] Optionally, the vehicle control device 500 further includes:

[0155] The updating unit 504 is configured to:

[0156] Determining an expected return function corresponding to the control parameter adjustment model;

[0157] Evaluating the control parameters output by the control parameter adjustment model using the expected reward function;

[0158] The control parameter adjustment model is updated according to the evaluation result.

[0159] Optionally, the adjusting unit 503 is further configured to:

[0160] Determining whether the second calculation control parameter reaches a preset parameter threshold;

[0161] If the second calculation control parameter reaches a preset parameter threshold, a prompt message is issued;

[0162] Adjust the control parameter according to the corresponding feedback information sent by the driver in response to the prompt message;

[0163] Adjust the target vehicle according to the adjusted control parameter.

[0164] Please refer to Figure 6 , which is a schematic structural diagram of the server of the present invention. The server 600 in this embodiment includes at least one processor 601, at least one network interface 604 or other user interfaces 603, a memory 605, and at least one communication bus 602. The server 600 optionally includes a user interface 603, including a display, a keyboard, or a pointing device. The memory 605 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory. The memory 605 stores execution instructions. When the server 600 runs, communication occurs between the processor 601 and the memory 605, and the processor 601 calls the instructions stored in the memory 605 to execute the above vehicle control method. The operating system 606 includes various programs for implementing various basic services and processing tasks according to the hardware.

[0165] The server provided by the embodiment of the present invention can execute the technical solutions of the embodiments of the above vehicle control method. The implementation principles and technical effects are similar, and will not be elaborated here.

[0166] The embodiment of the present invention also provides a computer-readable medium, including computer execution instructions, and the computer execution instructions can enable the server to execute the vehicle control method described in the above embodiments. The implementation principles and technical effects are similar, and will not be elaborated here.

[0167] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the above method embodiments; and the foregoing storage medium includes: ROM, RAM, magnetic disk, or optical disk and other media that can store program codes.

[0168] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A vehicle control method, characterized in that, It includes the following steps: Determine the first psychological data of the driver corresponding to the target vehicle, where the first psychological data includes emotion data and / or attention data; Input the first psychological data into a control parameter adjustment model to obtain a first calculated control parameter. The control parameter adjustment model is obtained by training a control parameter training sample, and the control parameter training sample includes the training psychological data corresponding to at least one driver and the control parameter corresponding to the training psychological data; Adjust the target vehicle based on the first calculated control parameter; Determine the second psychological data of the driver after adjusting the target vehicle based on the first calculated control parameter; If the second psychological data does not reach a preset adjustment threshold, determine a second calculated control parameter based on the second psychological data and the control parameter adjustment model; Adjust the target vehicle based on the second calculated control parameter; Judge whether the second calculated control parameter reaches a preset parameter threshold; If the second calculated control parameter reaches the preset parameter threshold, send a prompt message; Adjust the control parameter according to the corresponding feedback message sent by the driver for the prompt message; Adjust the target vehicle according to the adjusted control parameter.

2. The method according to claim 1, wherein The determination of the first psychological data of the driver corresponding to the target vehicle includes the following steps: Collect the first physical characteristic parameters of the driver; Input the first physical characteristic parameters into a driver recognition model to obtain the first psychological data.

3. The method according to claim 2, wherein The method further includes the following steps: Obtain the physical characteristic parameters corresponding to at least one driver and the actual psychological data corresponding to the physical characteristic parameters; Preprocess the physical characteristic parameters corresponding to the at least one driver and the actual psychological data corresponding to the physical characteristic parameters; Perform iterative operation according to the preprocessed data and the initial driver recognition model until a preset iterative termination condition is reached; Determine the initial driver recognition model when the preset iterative termination condition is reached as the driver recognition model.

4. The method according to claim 3, characterized in that, The method further includes the following steps: Judge whether the number of iterations reaches a preset value. If so, determine that the preset iterative termination condition is satisfied; Or, Judge whether the model parameters corresponding to the initial driver recognition model converge. If so, determine that the preset iterative termination condition is satisfied.

5. The method according to claim 1, wherein The method further includes the following steps: Determine the expected return function corresponding to the control parameter adjustment model; Evaluate the control parameter output by the control parameter adjustment model through the expected return function; Update the control parameter adjustment model according to the evaluation result.

6. The method according to claim 1, wherein The method further includes the following steps: Update the control parameter adjustment model through the second psychological data and the first calculated control parameter.

7. A vehicle control device, characterized in that, It includes: A determination unit for determining the first psychological data of the driver corresponding to the target vehicle, where the first psychological data includes emotion data and / or attention data; A calculation unit, configured to input the first psychological data into a control parameter adjustment model to obtain a first calculated control parameter, where the control parameter adjustment model is obtained by training a control parameter training sample, and the control parameter training sample includes training psychological data corresponding to at least one driver and control parameters corresponding to the training psychological data; An adjustment unit, configured to adjust the target vehicle based on the first calculated control parameter; The determining unit is further configured to: determine the second psychological data of the driver after adjusting the target vehicle based on the first calculated control parameter; if the second psychological data does not reach a preset adjustment threshold, determine a second calculated control parameter based on the second psychological data and the control parameter adjustment model, and adjust the target vehicle based on the second calculated control parameter; The adjustment unit is further configured to: determine whether the second calculated control parameter reaches a preset parameter threshold; if the second calculated control parameter reaches the preset parameter threshold, send a prompt message; adjust the control parameter according to the corresponding feedback message sent by the driver in response to the prompt message; and adjust the target vehicle according to the adjusted control parameter.

8. A computer device, characterized in that, Comprising: At least one connected processor, memory, and transceiver; Wherein, the memory is used to store program codes, and the processor is used to call the program codes in the memory to execute the steps of the vehicle control method according to any one of claims 1 to 6.

9. A computer storage medium, characterized in that, Comprising: Instructions that, when run on a computer, cause the computer to execute the vehicle control method according to any one of claims 1 to 6.

Images